Biometrics is the science of using unique personal and physical characteristics as a way to verify the claimed identity of a person. One of the oldest biometric sciences involves the use of fingerprints to identify a person. Since fingerprints are formed during fetal development, and are considered to remain unchanged throughout one's life, except for the rare cases where the fingerprint may change due to diseases or scars.
It is widely accepted that no fingers have identical prints, even from the same person or from identical twins. The systematic classification of fingerprints began in the 1800s and has been developed since that time through the extensive use of fingerprints by law enforcement and forensic societies.
In current usage, a fingerprint comparison is based on minutia (i.e., individual characteristics within the fingerprint pattern). For example, these minutia can include ridge endings and bifurcations in the fingerprint. The relative position of minutia is used for comparison, and according to empirical studies, two individuals typically will not have eight or more common minutia. A typical scan of a fingerprint contains 30-40 minutia.
Since fingerprints have a large amount of data, fingerprint biometrics have a high inherent accuracy that enables low rates of false acceptance (when a biometrics system wrongly verifies the identity of a person) and false rejection (when a biometrics system is unable to verify the legitimate claimed identity of a person).
In recent years, there has been much interest in developing fingerprint biometrics systems. For example, there are several commercially available fingerprint systems. One system is available from Veridicom of Santa Clara, Calif. A second system is available from IDEX AS of Heggedal, Norway.
An important aspect of any fingerprint biometrics system is the accurate capture of the fingerprint image. To ensure accurate capture, one approach taken by companies such as the Veridicom, is to have an array of sensors that covers the entire area of the fingerprint. One disadvantage of these full-array type of finger print imagers is that these imagers require a large sensor array to perform the image capture. This prior art fingerprint imager typically employs a sensor array whose area is commensurate with the area of the entire fingerprint. The large sensor array requires a sensor for each point of the fingerprint. For example, a standard sensor array has 400,000 sensors arranged as 500 pixels by 800 pixels so that the entire image of the fingerprint can be captured at one time.
It is well known that the cost of an integrated circuit is a function of the size or area of the integrated circuit. It would be desirable for there to be a mechanism to image a fingerprint by using an array of sensors that has fewer sensors than the number of sensors needed to cover the entire area of the finger print. By reducing the number of sensors needed to accurately capture the fingerprint image, the area of the integrated circuit can be decreased, thereby leading to cost savings. Unfortunately, the prior art approaches fail to provide such a mechanism.
A second approach followed by companies, such as IDEX AS of Heggedal, Norway, is to use a scan approach where a user is required to move his finger across an array or sensors that is substantially less than the total height of the finger. This approach is described in PCT patent application, International Publication Number WO 99/43258, entitled “Fingerprint Sensor.” Unfortunately, this scan-type imager has several drawbacks. First, the IDEX approach uses a curved sensor and sliding assemblies (see FIGS. 2 to 5) to adapt to and to guide fingers of differing sizes. These grooves and curvatures and assemblies increase the complexity and attendant costs to manufacture the fingerprint sensor. Second, the U-shaped groove can be uncomfortable for some and difficult to use by others. It is would be desirable to have a scan-type fingerprint imager with relatively planar sensors and without the assemblies described above.
PCT patent application, International Publication Number WO 98/58342, entitled “Method and Apparatus for Measuring Structures in a Fingerprint,” further describes a system in which a scan-type sensor can be utilized. This system employs a sensor array with two or more sensor lines in order to measure structures in the fingerprint smaller than the spacing of the sensors. Since this system uses two or more lines of sensors to image a single line of the fingerprint, the system must address the problem of combining these measurements of different lines into a single signal representing the single line in the fingerprint.
In order to combine these signals into a single signal, the patent application describes different correlation techniques that can be employed to combine signals measured from different lines of sensors. The system uses an adjustment scheme that corrects for the time delay between the signals from the sensors in different lines. Unfortunately, these techniques may increase the cost and complexity of the system. Consequently, it would be desirable to have a system that employs a single row of sensors that has comparable resolution and at the same time does not require complex signal correlation techniques.
Another problem faced by scan-type systems that are not encountered with full-array type sensor systems is that the movement of the finger needs to be determined. For example, to increase the accuracy of the captured image, the movement of the finger with respect to the sensor needs to be determined. Prior art systems perform this determination by using software executing on a host computer that is remote from the sensor. Unfortunately, the host computer solution may not adequately be able to handle the rate of data transfer between the sensor chip and the host computer unless a high-speed bus, which may be prohibitively expensive to implement, is utilized. Consequently, it is desirable for a system to have a single chip solution that has a sensor array and a mechanism for determining the movement of the finger.
Based on the foregoing, there remains a need for an apparatus and method for capturing an image of a finger print that employs a single sensor chip having an integrated navigation engine for determining the change in position of the finger with respect to time.